Projecting Semantic Information from a Language Independent Syntactic Model

• US 20090326924A1
• Filed: 12/15/2008
• Published: 12/31/2009
• Est. Priority Date: 06/27/2008
• Status: Abandoned Application

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Nguyen, Nghi H., Ziebell, Jonathan V., Crider, Anthony L.
• Application Number:
  US12/335,065
• Publication Number:
  US 20090326924A1
• Time in Patent Office: Days
• US Class Current: 704/9
• CPC Class Codes: G06F 40/30 Semantic analysis
• CPC Class Codes: G06F 8/30 Creation or generation of source code

Abstract

Embodiments for the conversion of Computational Independent Model (CIM) rule expressions into semantically non-ambiguous syntax trees are disclosed. In accordance with one embodiment, a method includes analyzing a sentential structure of a Computational Independent Model (CIM) rule expression for clauses. The clauses include at least one expression and at least one rule. The method further includes constructing a semantically non-ambiguous LF syntax tree from the CIM rule expression. The construction being implemented using a logical form (LF) model.

Claims

1. A method, comprising:

creating a logical form (LF) model that provides a semantically non-ambiguous LF syntax tree from a rule expression, the semantically non-ambiguous LF syntax tree representing one or more anaphoric references in the rule expression;

analyzing a sentential structure of a rule expression for a plurality of clauses that include at least one expression and at least one rule using the LF model, the plurality of clauses including a plurality of conditional clauses;

translating the rule expression into a semantically non-ambiguous LF syntax tree based on the LF model, the translating includes adjoining the plurality of conditional clauses by resolving one or more anaphoric references; and

presenting the semantically non-ambiguous LF syntax tree.

2. The method of claim 1, wherein the rule expression includes a Computational Independent Model (CIM) rule expression.

3. The method of claim 1, wherein the semantically non-ambiguous LF syntax tree includes at least one phrase marker, the at least one phrase marker being one of a sentential phrase marker or a projection phrase marker.

4. The method of claim 1, wherein the semantically non-ambiguous LF syntax tree includes a plurality of projections, and wherein at least two of the projections are linked by a variable, and each projection including one or more tuples.

5. The method of claim 1, wherein the analyzing the rule expression includes recursively descend through the sentential structure of the rule expression.

6. The method of claim 1, wherein the plurality of clauses further includes one of an event clause, a given clause, and a main clause.

7. The method of claim 1, wherein the adjoining the plurality of the conditional clauses includes using an if phrase operator to joining the coordinate phrase markers of two conditional clauses, and reversing an order of two conditional clauses

8. A computer readable medium storing computer-executable instructions that, when executed, cause one or more processors to perform acts comprising:

analyzing a sentential structure of a Computational Independent Model (CIM) rule expression for a plurality of clauses that include at least one expression and at least one rule;

constructing a semantically non-ambiguous LF syntax tree from the CIM rule expression using a logical form (LF) model; and

presenting the semantically non-ambiguous LF syntax tree.

9. The computer readable medium of claim 8, wherein the at least one rule includes one of an assertion rule, a constraint rule, a declaration rule, and a default rule, and wherein the at least one expression includes one of an event expression, a role expression, a value expression, a fact expression, and sentence expression.

10. The computer readable medium of claim 9, wherein each of the assertion rule, default rule, and declaration rule includes at least one of an event clause, a given clause, a main clause, or a conditional clause, and the constructing includes creating a phrase marker for one of an event clause, a given clause, a main clause, or a conditional clause.

11. The computer readable medium of claim 9, wherein the constructing includes creating a modal phrase marker for one of a permission modal operator, an obligation modal operator, or a prohibition modal operator of the constraint rule.

12. The computer readable medium of claim 8, wherein the plurality of rules includes at least one rule that comprises multiple clauses, and wherein the providing includes adjoining the multiple clauses into coordinate phrase markers.

13. The computer readable medium of claim 12, where adjoining the multiple clauses includes using an upon phrase operator to join the coordinate phrase markers of multiple event clauses, and using a given phrase operator to join the coordinate phrase markers of multiple given clauses.

14. The computer readable medium of claim 12, wherein adjoining the multiple clauses includes using an if phrase operator to joining the coordinate phrase markers of two conditional clauses, and reversing an order of two conditional clauses.

15. The computer readable medium of claim 9, wherein the constructing includes creating an event phrase marker that projects an event projection for the event expression.

16. The computer readable medium of claim 9, wherein the constructing includes consuming role expression to create tuples for the fact expression to range over.

17. The computer readable medium of claim 9, wherein the constructing includes projecting a fact type of the sentence expression as a propositional phrase marker and projecting an n-tuple of variables.

18. The computer readable medium of claim 9, wherein the value expression includes one of a quantifier expression, a literal expression, an instance expression, a mathematical expression, or a fact value expression, and wherein the constructing includes one of project a literal expression node for the literal expression that represents a literal value and binding the literal value to a variable, project a constant instance that represents a name value and binding the name value to the variable, decomposing the mathematical expression into an expression tree that surfaces the mathematical computation as an expression value and binding the expression value to the variable, or decomposing a quantifier expression by determining a relevant quantifier type and projecting a quantifier as one of cardinal or universal, and binding the quantifier to the variable.

20. The transformer of claim 18, wherein the semantically non-ambiguous LF syntax tree includes at least one of one or more phrase markers, one or more projections, one or more quantifiers, one or more variables, and one or more tuples.

19. A transformer, comprising:

an input component that receives a Computational Independent Model (CIM) rule expression, the CIM rule expression including at least one anaphoric reference, the at least one anaphoric reference being a pronominal link between an antecedent reference and one of a definite description or a pronoun;

a logical form (LF) model that resolves the at least one anaphoric reference into a semantically disambiguated representation;

an algorithm that generates a semantically non-ambiguous LF syntax tree from the CIM rule expression based on the LF model, the semantically non-ambiguous LF syntax tree including the semantically disambiguated representation; and

an output component that present the semantically non-ambiguous LF syntax tree.